16:16:02 <djmitche> #startmeeting weekly meeting
16:16:02 <bb-supy`> Meeting started Tue Jul  7 16:16:02 2015 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:16:02 <bb-supy`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:16:02 <bb-supy`> The meeting name has been set to 'weekly_meeting'
16:16:08 <djmitche> sorry, geez
16:16:53 <sa2ajj> topics for today?
16:16:57 <djmitche> so first up is nine
16:17:03 <djmitche> #topic status of nine beta
16:17:16 <tardyp> not to much to say on my side
16:17:20 <djmitche> I'm a behind on reading mail, but I've seen bugs filed and some ML conversations about it
16:17:26 <tardyp> was pretty busy for works this week
16:17:39 <sa2ajj> i'm exploring CentOS :)
16:17:40 <tardyp> I have a big authz PR for review https://github.com/buildbot/buildbot/pull/1730
16:18:01 <sa2ajj> i'll have a look
16:18:21 <djmitche> oh, great, auth will help!
16:18:23 <tardyp> last week, we had _VV do a bunch of tests and comments
16:18:31 <sa2ajj> looks very interesting
16:18:32 <tardyp> jared did start a migration guide
16:18:32 <djmitche> #info Pierre has a PR up to add authentication support
16:18:42 <tardyp> authorization
16:18:48 * sa2ajj nods
16:18:53 <djmitche> #info sorry, authorization
16:18:54 * sa2ajj to the latter
16:19:07 <djmitche> #info Vitali did some testing and commenting
16:19:08 <tardyp> authentication has been there for a year
16:19:15 <djmitche> #info jared working on migration guide
16:19:29 <djmitche> so no huge roadblocks -- just things to do/fix?
16:19:47 <sa2ajj> jaredgrubb: is the work-in-progress available anywhere?
16:19:57 * sa2ajj did not follow PRs that actively this week...
16:20:14 <jaredgrubb> there is a PR for it … and tardyp & i did have some questions for you :)
16:20:37 <sa2ajj> right, then that'd be an AP for me: read, understand and answer :)
16:20:39 <jaredgrubb> there's not much to it yet, but i'm just filling it out as i go … and hoping others do as well … i'll try to merge in vitali's notes as well
16:20:39 <djmitche> https://github.com/buildbot/buildbot/pull/1756
16:21:00 <djmitche> #action sa2ajj to look at https://github.com/buildbot/buildbot/pull/1756 and answer questions there
16:21:08 * sa2ajj nods
16:21:27 <tardyp> question is about doc refering to import xxx from plugins
16:21:41 <sa2ajj> let's continue in the PR
16:21:48 <tardyp> ok
16:22:04 <djmitche> ok, sounds good
16:22:09 <djmitche> #topic donation buttons
16:22:18 <djmitche> verm__: you around for an update?
16:22:34 <djmitche> (agenda is from https://etherpad.mozilla.org/buildbot-agenda btw)
16:22:47 <djmitche> we can loop back
16:22:51 <djmitche> #topic GSoC updates
16:23:01 <djmitche> first up - stats (prasoon)?
16:23:28 * sa2ajj did not follow that
16:24:01 <stibb> Yes
16:24:03 <tardyp> stibb: ?
16:24:22 <stibb> We think final pl to be ready tomorrow
16:24:35 <djmitche> what sorts of changes does it contain?
16:24:43 <stibb> I hope he ll have lot of reviews
16:24:54 <stibb> Rework after the review by Pierre
16:25:00 <stibb> So mq
16:25:16 <djmitche> great
16:25:22 <sa2ajj> stibb: please include '@sa2ajj' so i would not miss that...
16:25:34 <djmitche> #info Prasoon hoping to have a reworked pull request up tomorrow
16:25:42 <djmitche> material design (chase)?
16:25:43 <sa2ajj> include = mention in comments :)
16:25:47 <stibb> And user can define their own capture method
16:25:50 <stibb> Ok
16:25:54 <sa2ajj> thx
16:26:00 <djmitche> stibb: great, so they can capture their own stats
16:26:01 <tardyp> I think we really should start merging something
16:26:07 <stibb> Yes
16:26:11 <tardyp> because the branch start to diverge much
16:26:25 <tardyp> from my side
16:26:27 <stibb> And starting thinking about next half of the project
16:26:35 <stibb> Also their is doc that is started
16:26:46 <tardyp> andras and chaze are going well
16:26:53 <tardyp> the easy stuff are done
16:27:01 <tardyp> now the hard stuff is starting.
16:27:15 <sa2ajj> tardyp: can you elaborate?
16:27:22 <tardyp> for andras, he is strugling with the multi tab optimization
16:27:33 <tardyp> trying to have only one tab doing the request for the others
16:27:40 <tardyp> and maintaining a shared cache
16:27:52 <djmitche> I imagine that runs into some browser compatibility issues?
16:28:04 <tardyp> not really
16:28:21 <tardyp> browser compatiblity is not something there are so much nowadays
16:28:27 <tardyp> it is more raceconditions
16:28:32 <sa2ajj> ooh
16:28:46 <tardyp> there is the election of a master tab which owns the websocket
16:28:51 <djmitche> oh, wow
16:29:03 <tardyp> but in the current design, each tab does the data query itself
16:29:09 <tardyp> so this is very racy
16:29:22 <djmitche> #info andras and chase (material design and new data API) are into the "hard stuff"
16:29:26 <tardyp> so I suggested to have a cache system, so that only one tab maintain the data
16:29:42 <tardyp> and for chase, he will be starting the buildpage this week
16:29:47 <djmitche> #info andras working on sharing a websocket between multiple open tabs; running into race conditions
16:29:55 <djmitche> great
16:30:02 <tardyp> and then there will be the force scheduler ui
16:30:14 <djmitche> I had a peek at the demo - it looks great and the interface feels very natural
16:30:14 <tardyp> those are the hardest part of the UI, probably
16:30:18 <djmitche> indeed
16:31:00 <tardyp> people are worrying about this new UI
16:31:03 <djmitche> oh?
16:31:25 <tardyp> they are starting to write plugin for the bootstrap UI, and wondering if these will be obsoleted by the md ui
16:31:47 <tardyp> for me the boostrap UI shall continue to be maintained, and the md UI is for now more like an alternative
16:32:04 <djmitche> that seems like a big maintenance burden
16:32:08 <tardyp> and a demoinstration on how it is easy to create a pure customized ui
16:32:12 <djmitche> ah
16:32:21 <djmitche> so we won't necessarily maintain the md UI?
16:32:37 <tardyp> If people care about it yes
16:32:56 <tardyp> but I doubt that we will be able to have the same level of functionality in one gsoc
16:33:04 <djmitche> true
16:33:33 <djmitche> #info bootstrap-based UI will remain the default for the forseeable future, so it's worth building plugins against it
16:33:54 <djmitche> sa2ajj: varun: how are things looking?
16:33:58 <tardyp> for now the md ui does not event support plugis
16:34:19 <djmitche> last I heard, work was afoot on the client script
16:34:26 <sa2ajj> i did not really check the changes :(  i'm going to suggest to file a PR, this way it gets better review...
16:34:36 <varun> Great :) Gotta finish my project too
16:34:37 <djmitche> yeah, good idea
16:35:10 <djmitche> It might be nice to get the client working *without* sending patches first
16:35:18 <sa2ajj> varun: please try to split it in independent PRs
16:35:20 <djmitche> then in a second PR, get it to send patches (and the force scheduler to accept them!)
16:35:51 <sa2ajj> if functionality is not "really working", provide some instructions on how to use it.
16:35:57 <sa2ajj> varun: ^^
16:36:08 <djmitche> #info try-via-force (varun) broken up into smaller units, hopefully moving to PR's soon
16:36:10 <varun> Yeah, was having trouble finding a way to use mandatry arguments
16:36:30 <varun> I tried decoding harcoded get request, it worked
16:36:58 <sa2ajj> varun: we can have a look at this tomorrow at about the same time, is it ok?
16:37:08 <varun> yeah, sure
16:37:14 <sa2ajj> good, thank you
16:37:15 <djmitche> great!
16:37:22 <djmitche> and finally, Lurimax and jaredgrubb
16:37:30 <djmitche> and the Python 3 port
16:37:57 <sa2ajj> we still need to add a Buildbot slave to perform Py3 check
16:38:14 <sa2ajj> so our buildslave part is not diverting... :)
16:38:37 <jaredgrubb> jørn is doing well .. buildslave is pretty much done and he's starting on the master stuff
16:38:43 <sa2ajj> and, hopefully, the master part would be OK as well when Lurimax is done
16:38:45 <jaredgrubb> he's kinda stalled on some of the twisted parts
16:38:50 <djmitche> I was going to say the same -- some automated way of preventing backsliding
16:39:00 <djmitche> even if htat's greps for 'except' not followed by 'as', etc.
16:39:13 <dwlocks> any opinions on something like  def main(argv):    in a script?
16:39:33 * djmitche waves
16:39:35 <sa2ajj> dwlocks: please add a context :)
16:39:43 <bobhood> Let me point out that the twisted issues were known going in, so that risk is still nominal at this point.
16:39:54 <dwlocks> sa2ajj, there isn't much.
16:40:02 <djmitche> bobhood: absolutely
16:40:04 <sa2ajj> djmitche: please add an AP for me to review that ticket and [try to] add the proper builder to metabb
16:40:15 <dwlocks> if __name__ == __main__:
16:40:29 <dwlocks> main(sys.argv)
16:40:31 <jaredgrubb> @bobhood: yes agreed … right now he's run into a bytes vs str issue and it's not clear how to solve that in the way that twistd will want it solved (in buildbot, we'd use python-future, but twisted doesnt do that … yet?)
16:40:36 <djmitche> #action sa2ajj to look at #3287 regarding py3 compatibility testing
16:40:41 <sa2ajj> dwlocks: do not forget '' around __main__ :)
16:40:51 <dwlocks> as opposed to the more usual
16:40:59 <dwlocks> main()
16:41:11 <dwlocks> sa2ajj, right.  quick typing.
16:41:14 <djmitche> dwlocks: I like passing argv in since it makes it much easier to test main without setting sys.argv
16:41:16 <jaredgrubb> (i shouldnt say twisted "doesnt want to" … i should say, we are not sure yet how to proceed… they havent yet chimed in)
16:41:37 <djmitche> OK, lmk if you want me to try to rouse some people from slumber
16:41:38 <dwlocks> djmitche, that's what guido says, but it breaks setup.py
16:41:50 <djmitche> oh, I wouldn't do that in setup.py
16:42:00 <djmitche> mind deferring this for a moment while we wrap up the meeting?
16:42:01 <dwlocks> nono.  not in setup.py itself.
16:42:08 <dwlocks> sure.  sry
16:42:11 <djmitche> we may be done, in fact -- any other business?
16:42:32 <tardyp> nope
16:42:35 <djmitche> #info python 3 port: buildslave work largely finished, but needs testing to ensure it doesn't regress (#3287)
16:42:49 <djmitche> #info python 3 port: master work is more intertwined with twisted
16:43:04 <djmitche> #topic GSoC Mentor Summit
16:43:12 <djmitche> we need to decide who we're sending to the summit
16:43:16 <sa2ajj> i think the approach i offered in the ticket (still need to review that) would show us problems w/ Py3 and not twisted
16:43:18 <djmitche> and to start -- who of the GSoC folks *want* to go?
16:43:30 <tardyp> I pass
16:43:45 <bobhood> Ditto
16:43:47 <sa2ajj> djmitche: i should go to US some time soon and i still did not apply for a visa :(
16:43:59 <djmitche> I can loan you a mastercard
16:44:07 <sa2ajj> :)
16:44:08 <djmitche> I don't have American Express, sorry
16:44:10 <jaredgrubb> (note google does pay travel and lodging)
16:44:20 <jaredgrubb> (in case that sways anyone :)
16:44:24 <sa2ajj> visa as in being able to get to the country :D
16:44:52 <djmitche> (we can send two delegates)
16:44:54 <stibb> Not against
16:44:55 <djmitche> I know just teasing
16:45:11 <djmitche> I'd like to go, having skipped last year
16:45:15 * sa2ajj feels deeply teased..
16:45:18 <jaredgrubb> i would go, but i went the last two years (last year as a lottery winner) … so i'll add my name to the end of the list
16:45:34 <djmitche> ok, I'll put the word out to the entire list of mentors (at least one isn't in irc) and figure it out
16:45:47 <sa2ajj> sounds good
16:46:02 <djmitche> #action dustin to poll mentors to decide who will go to the summit
16:46:12 <djmitche> does that wrap it up?
16:46:29 <jaredgrubb> nothing more from me
16:46:34 <djmitche> again I'm really sorry to be 15 minutes late, what a jerk :(
16:46:35 <sa2ajj> sounds like it
16:46:38 <djmitche> #endmeeting